State of the art
Data analytics is an emerging tool for law enforcement. Police already rely widely on gathering and analysing data digitally. For example, most departments havean electronic system for recording an officer’s notes or calls to a tip line. Often, critical data such as the time, place, type of crime, and witness details are captured (Nath. 2006). These notes are often collections of data that a police officer uses to draw inferences about an offender. Police commonly have a system in place to record this information electronically. Many departments use these RMS only as a way of archiving data. However, police are becoming increasingly cognisant of the analytic opportunities such data enable.
Data analytics can take many forms: entity extraction, clustering techniques, rule mining, classification, and social network analysis, for instance. Clustering can be used to find similarities among groups of criminals, groups of crime occurrences, and times and locations where incidents happen. Police departments that use clustering techniques can ‘identify suspects who commit crimes in similar ways or distinguish among groups belonging to different gangs’ (Chen et al., 2004, p. 51), generalising the modus operand! files that police have used for more than a century. Police departments can also develop predictive models that suggest locations, times, or situations that have elevated risks for crime occurrence. The use of data analytic techniques has two principal advantages that complement the experience of officers: they demonstrably avoid bias (both social and cognitive), and they are measurably more accurate, increasing effectiveness. Sometimes, data analytics can identify patterns of which police were unaware. In its strong form, this has precipitated a shift toward what has commonly become known as ‘predictive policing’. In a weaker form, knowing these patterns can suggest new investigative directions that are likely to be productive.
Handling incident data within an analytic framework also helps make connections between cunent and previous crime patterns which may be known to officers but are easily missed because of shift turnover (in the short term) and employee turnover (in the medium tenu). A passive RMS system preserves this information, but officers must actively look for it; an analytic system can ‘push’ potentially useful patterns to officers.
Hot spot policing
Geography is also a critical component of many patterns of crime. A method known as ‘hot spot’ policing allows law enforcement to identify where offences tend to be concentrated. Hot spots are identified by plotting the historical patterns of previous incidents. When these plots show clusters - geographically close incidents - these are called hot spots (Nath, 2006). They suggest locations where proactive deployment can suppress potential crimes. Hot spot policing is defined as the practice of ‘policing focused on small geographic places or areas where crime in concentrated' (Koper, 2014, p. 123). Identifying areas of crime through computerised aggregates of data dates back to the 1970s (McEwen & Taxman, 1995). Hot spot policing has since emerged as one of the most critical innovations in the policing (Koper, 2014).
Information on hot spots allows police to optimise the geographic allocation of scarce resources and skills. In an effort to reinforce the ‘policing by consent’
Leveraging police incident data 57 approach, a hot spot might benefit from greater police visibility and the form that visibility should take, such as police interaction with civilians rather than patrol. When Bill Bratton (2011) was chief of the Los Angeles Police Department, using a data-enabled approach to policing reduced crime in certain areas merely by having police engage more with the community. The presence of competent, sophisticated community-oriented policing heightened the community’s confidence that offenders would be apprehended and held responsible for then crimes. Concentrating officers in areas identified as having higher incidence of crime than others (also known as ‘cops on the dots’), Los Angeles experienced a consistent reduction in crime over seven consecutive years. Rather than being merely reactive, police can concentrate their analytical and diplomatic efforts on resolving underlying issues proactively in a hot spot for which crime may merely be a symptom (Кореї; 2014).
The sites of burglaries, for example, are more likely to be burgled again. Ber-nasco found that scholars such as Johnson et al. (2007) concluded that ‘in the wake of a domestic burglary, not only the property itself but also properties near the victimised property have an elevated burglary risk’ (Bemasco, 2008, p. 412). The risk of a repeat burglary can remain heightened for at least a month. If burgled homes are at an increased risk of re-victimisation, then police can ‘harden’ the target by increasing their presence around such properties (Bowers & Johnson, 2004). Hardening the target can also include identifying individual characteristics that make the dwelling or the victim susceptible to theft. Research indicates that the same offenders are likely to burgle the same dwelling again and often with the same modus operandi (Bemasco, 2008).
Less obviously, researchers have shown that the social networks of, and around, criminals have their own hot spots. The distribution of gunshot victims is much more likely in the social network of those who have been arrested, and even within such social networks is highly skewed (Papachristos et al., 2015). If police can identify those regions of criminal social networks that have the highest risk of violence, they can begin to be proactive about pre-empting such violence.
Hot spot policing has thus established itself an innovative approach to identifying areas with substantial crime so resources can be allocated accordingly. This approach is predicated on the assumption that crime is not randomly distributed. Statistics indicate that areas that have experienced crime are more likely to be the location of future crime. Studies across US cities have shown that 5 percent of addresses tend to be associated with half of a city’s crime (Кореї; 2014).
Intelligence is key to reducing crime successfully among repeat offenders. It is well established that a small portion of the population is disproportion-ally responsible for the majority of crime (e.g., Wolfgang, Figlio, & Sellin, 1972). Versteegh, Van Der Plas, and Nieuwstraten (2013) found that in 2002 ‘85% of these veiy active repeat offenders still reoffended in the same year. In 2006, that share dropped to 70%, after which the share of reoffending repeat offenders rose slightly to 77% in 2008’ (p. 72). Intelligence helps identify prolific offenders and promotes activities that will deter them from offending.
For those who are not deterred, vigilance heightens the likelihood of their arrest should they offend (Groff et al., 2015). In a controlled experiment by the Philadelphia Police Department that targeted prolific offenders, locations with an offender-focused approach saw a 50 percent reduction in violent felonies (Groff et al., 2015).
Hot spot policing thus has a track-record of reducing crime. Bratton (2011) squarely attributes part of that reduction to the role of hot spots play in predictive policing. Adding more police to a high-crime area deters crime: ‘with no criminal act, there was no arrest, no conviction, and no incarceration of perpetrators’ (p. 66). Also, the community feels more secure, knowing that perpetrators are more likely to be identified and caught. Statistically, other cities that practise hot spot policing also saw a reduction in crime, sometimes by 80 percent (Braga, Papachristos, & Hureau. 2014).
Larger police departments usually have greater discretion to allocate resources and analytical tools to hot spot policing. A study of hot spot policing across 198 agencies found that the practice is employed disproportionately by larger jurisdictions (Koper, 2014). Departments that did not participate tended to be smaller, with fewer than 100,000 residents. They may not have capacity, but they may also feel that there are diminishing returns on hot spot policing in smaller communities. We deliberately picked a smaller conununity for this exploratory study, in part, to investigate this.
